When countries manipulate economic data

Conspiracy theorists have a field day whenever official statistics look conveniently better just before elections. Whether statistics are really manipulated for political gain is hopefully less frequent than what they assume. We know it is currently done in Argentina, was done by the Soviets and their allies, and used to be done by some countries to qualify as poor in the United Nations' eyes. Those cases were rather obvious, but how could you recognize the more subtle ones?

Tomasz Michalski and Gilles Stoltz use Benford's Law, the distribution of first digits in economic figures, to determine the likelihood of manipulation across a large set of countries. While this does not catch a country red-handed, it gives probabilities, and one can analyze this against a set of indicators to determine what would drive them to cheat. Michalski and Stoltz find that higher likelihood of cheating is associated with fixed exchange rates, high negative assets, negative current accounts or subject to capital flow reversals. So it seems that this kind of cheating is not for internal consumption, but rather to deceive international financial markets. The authors did the analysis on balance of payments data, though, so the picture may be quite different when looking at unemployment, GDP or inflation data.

Early uses of accounting: to help in firm management or to pursue an agenda?

You may think that accounting practices are straightforward and have been in place for a long time. Actually, good practices are actually fairly recent, especially in terms of making them useful diagnostic tools for firm management. But with sophistication comes also the temptation to become creative and use accounting for purposes that are borderline legal, such as escaping taxation, or outright fraud. For this, you would need to be a sophisticated accountant, and one would think that one would not find such sophistication a century ago, let alone during the British Industrial Revolution.

Steven Toms and Alice Shepherd show that in the second case there were surprising sophistication, with creative accounting being used by industrialists to counter the "Ten-Hour" movement that sought to limit work hours. Specifically, they show how the the numbers from a cotton manufacturer were used in the policy debate and how his creative accounting made it appear as though he was facing excruciatingly high fix costs and thus low profits. Where he got creative is with the treatment of capital accumulation, thereby proving that the accusation of making most of his supposedly high profits during the last hour of the shifts was not true.

Meta-analysis of the elasticity of intertemporal substitution

The elasticity of intertemporal substitution is one of the most estimated parameters in Economics. Why is it estimated over and over again? Because some results are positive, some are negative and some are zero. To have a clearer idea of what its true value is, we have to keep estimating it. However, the econometricians also need to get their results published, and the publishing tournament has not only an impact on which results get published but also on which ones the econometricians submit for publication.

Tomáš Havránek performs a meta-analysis of estimates of the elasticity of intertemporal substitution. That is, he gathers 169 studies and looks at their 2735 estimates. He finds significant under-reporting of results close to zero or negative, because of this publication bias. While the published mean is 0.5, the true mean should somewhere at 0.3 to 0.4. Negative results make little sense, but they can happen with some draw of the data. If editors and referees systematically discard such results, and positive ones, no matter how large they are, get a pass, we have a bias. But given the distribution of published ones, and knowing this bias, one can infer the full distribution of estimates, and hence Havránek's new estimates.

Identifying monetary policy "shocks"

I have always found the empirical monetary policy literature rather frustrating. It is entirely based on the premise that one can identify monetary policy shocks. First, I am not sure what is really meant by a shock. Is it any change in a policy variable? Not changing it may be a surprise, as we recently witnessed by with the recent FOMC decision not to throttle quantitative easing. And how much a change is anticipated matters as well. The recent emphasis on forward guidance makes the interpretation of an interest change very different from the surprise actions from a few years ago. Second, the empirical identification of those shocks seems doubtful at best. Either you take a VAR and interpret residuals as shocks (never mind those will be significantly different across specifications), or you try to quantify some narrative of policy decisions, sorting out rather subjectively what was a surprise and what was expected. Third, a monetary policy shock should be measured differently under different policy regimes. There is no point on focusing on the Federal funds rates (or a Taylor rule) when the policy focuses on the money supply, for example.

The reason for this rant is that I came across a paper by Martin Kliem and Alexander Kriwoluzky who try to reconcile the VAR and narrative approaches, which of course is impossible. What they highlight though is that both are fraught with error. They find this by plugging the narrative measure into a VAR and they conclude that there is measurement error in the narrative measure and misspecification error in the VAR. That should surprise no one, but needs to be pointed up, with so many people relying blindly on these instruments.

Memorable goods

In macroeconomics, one distinguishes between non-durable and durable consumption goods. This distinction is important, as the cyclical nature of the two is very different. Durables are very volatile, as households like to postpone their acquisition in recessions. Non-durables are extremely smooth, however. The later is what most models have in mind when thinking about consumption, while the first are more like investment goods, but at the household level.

Rong Hai, Dirk Krueger and Andrew Postlewaite think we should add a third category: memorable goods. These are non-durable goods that may not last long physically, but we keep good memories about them and thus they continue to provide utility in the future. In essence they are also durable goods, but they are not counted as such in national accounting. Some examples the authors provide are Christmas gifts whose memories last through the year. The same applies to vacations, going out, clothes, and jewelry. Using the consumption expenditure survey, the authors find that memorable goods lie somewhere between durables and non-durables in terms of cyclical properties. As they account for about 14% of outlays, their presence matters quantitatively. In fact, they can fully explain some observed deviations from the permanent income hypothesis. A paper to remember and cherish for a long time.

Can we measure smoking behavior?

In any survey, we must consider the issue of imperfect recall or misrepresentation in self-reported assessments. People do not remember precisely how much they spent on this or that, and they they may not recall how long they have been unemployed. For some questions, social pressure may also be a factor. For example, one may not concede on using illegal drugs or one may misreport smoking behavior. The extend of such biases can be measured though, if one has access to administrative data or some other objective measure. The results are often disappointing (Examples discussed here: 1, 2).

Vidhura Tennekoon and Robert Rosenman criticize the fact that the measures against which surveys responses are compared are taken as perfect gold standards. Specifically they look at the biochemical assessment of smoking status, whose results the literature never doubts and which makes self-reported smoking status look really unreliable. Once you use statistical methods that concede that the biochemical assessment may also include some measurement error, they realize that it may be just as bad as the self-assessment. One can thus not exclude that the self-assessment may actually be a better indicator that an independently recorded measure.

Travel time and the border effect

The border effect describes a striking feature of the data on trade volumes. Volumes typically decrease with distance traveled, with a jump down when a border has to be crossed. The size of this effect is mostly estimated with distance data taken from straight lines between trading areas, and often by simply taking the center of those regions. With considerable work, one can do better.

Henrik Braconier and Mauro Pisu determine the distance along roads as well as travel time for within-Europe trade, and this for almost 50,000 city pairs. While this neglects cargo train traffic, which has a substantial share of international traffic in Europe, this is as precise as it can get, I suppose. The interesting bit is that once a border is involved, travel distance and time are about 10% longer for town pairs that are equidistant when measured as a straight line. That means that literature has over-estimated the border effect by about as much. One could, however, also argue that the border effect is precisely stemming from the fact that it leads to travel time losses, at least in part, and that there is therefore no over-estimation. It depends what you really mean by border effect.

Why is living in poor countries so cheap?

Theory tells us the law of one price should hold: the same good should have the same price throughout the world after taking into account exchange rates (and transportation costs). Yet, there is plenty of empirical evidence that this is not true. And anecdotal evidence, too, think of how it is noticeably less expensive to live in developing countries. Why?

Daniel Murphy offers a new and simple explanation: complementarity between tradable and non-tradable goods. In a rich country, more non-tradable goods are available as complements, thus providers of tradable goods can charge higher prices if competition is not perfect. This conjecture is supported by empirical evidence showing that where more complements are used the prices of tradables are also higher. But given that the extend of complementarity seems to change from one country to the other, it seems to me that we are not really talking about the same good. We may measure it as the same good, but people seem to perceive it as a different good. It is just a measurement issue.

Biased taxable income elasticities

Anytime you apply a distortionary tax, it bring well-being losses from the distortion (although the revenue can be used for well-being enhancing public goods). In addition, there are social losses that arise from the fact that people try to evade the tax by shift to other goods, go informal, or in the case of income shift compensation to non-taxable benefits or other amenities like more flexible work hours. Traditionally, the literature has evaluated the deadweight loss from taxation by looking at the income elasticity of the tax. That may be too simple a statistic in this case.

Brendan Epstein and Ryan Nunn show that ignoring the endogeneity of the non-taxed benefits and amenities leads to serious biases in the income elasticity and thus deadweight loss, to the point that it provide not good guidance on how to set tax rates. They basically do this by providing examples: build simple models, calibrate them, generate data from them and show that the usual empirical method provide crassly wrong estimates. An econometrician could in principle do better by taking all this in account, unfortunately data will be very hard to come by for this.

How much money laundering is there in Italy?

It is well known that the underground economy in Italy is substantial, and that an important share of this is due to illegal activity. Hence, there should be an important amount of money laundering going on, an amount that seems to be impossible to measure given that these activities precisely try not to get detected. But economists can be resourceful and try to pull it off, for example à la Steve Levitt.

Guerino Ardizzi, Carmelo Petraglia, Massimilano Piacenza, Friedrich Schneider and Gilberto Turati try to pull that off, reasoning that money laundering is performed by depositing cash, and that if there are more cash deposits in financial institutions of an Italian province where there is more activity from illegal syndicates, one should be able to back out how much of these deposits are due to money laundering. Concretely, they regress across provinces over four years cash deposits on a few controls, the number of detected extortion crimes and the number of drug dealing, prostitution and possession of stolen goods. One may have some qualms in using detected crimes, which may be a very poor proxy for actual crime, especially for a country that is so corrupt, but I suppose this is all we have. However, this regression assumes that those illegal syndicates stay within the confines of their province when they deposit their proceeds. Given the size of an Italian province (median inhabitants: 375,000), that seems like a real stretch. I guess we still do not know how much money laundering is going on in Italy.

Could obesity rates be even worse than expected?

The rise of obesity rates is now being called an epidemic, in particular in Anglo-Saxon countries. The fact that such a large portion of the population is now considered obese is quite alarming, considering that this was minimal a generation ago, and that the proportion of obese children is even in the double digits in some countries is mind boggling. Could it be even worse?

Yes, according to David Madden who claims that most obesity statistics are based on self-reports for weight and height, from which the BMI (body-mass index) is calculated. Under current standards, a BMI of 30 is considered obese. He suggests that a threshold of as low as 26 should be used to account for the reporting bias in weight. As this bias seems to increase over time (at least in Ireland), the threshold could move down even further. Obviously, this bias will depend on the environment (local culture, context of survey, for example) and could make correct measurement very uncertain. I guess the best way is to actually measure people. One should look into that.

Leaning against publication bias: about the experiments that do not work out

It is quite obvious that journals will only publish results that are significant in the statistical sense. If it turns out that X does not influence Y, in most circumstances there is little interest from editors and referees. Yet, this could be valuable, especially when this was rather costly to do. You want to avoid having someone else waste resources trying to do the same thing. And nowhere is this more important than in the outrageously costly experimental literature. (Yes I realize results could still be published as a working paper, but few people write working papers without the initial intention to publish).

Francisco Campos, Aidan Coville, Ana Fernandes, Markus Goldstein and David McKenzie report on seven such failed experiments, where it was not simply the lack of significant results that was causing trouble, but rather that they never came to the evaluation phase. While no designer of experiments will ever cite this paper, everyone should read it. Why did these experiments fail? Delays, hindrances from politicians, and low participation. The authors think this occurred because of some idealized vision of the experiment by the designers that cannot be translated into the field. In particular, one needs to cope with the local political economy, the incentive of field staff, and be a bit pragmatic in terms of program eligibility criteria. And in the end, if the experiment is not properly randomized, one cannot draw conclusions from it.

I wonder how many millions have been wasted on failed experiments that nobody hears about and nobody can learn from. And that is without counting the millions wasted on experiments that are impossible to generalize, and thus not useful even if "successful."

How randomized experiments can go very wrong

Randomized experiments are all the rage in some circles, for example labor economics and especially development economics. The principle is simple: create some intervention in some market, randomly draw a group of economic agents that has access to the intervention, leave the others out, compare outcomes. In all that, you hope the behavior of the non-participants is not affected by the presence of the program to the others. This can be a heroic assumption, for example because market prices may respond for everyone to the intervention.

Pieter Gautier, Paul Muller, Bas van der Klaauw, Michael Rosholm and Michael Svarer show an example where this assumption was violated. The intention was to see how helping Danish unemployed workers find jobs through enhanced guidance was successful. Those who were non-selected had to deal with the job search as usual. In that case, there were some regions where the experiment was not conducted but data still collected. In the two counties where the experiment was conducted, the number of vacancies markedly increased, which logically leads the treated and untreated to have a better shot at finding a job. But, of course, there is also a congestion effect: for the same number of vacancies, if some workers are getting better probabilities for finding jobs, it is getting worse for the others. In the Danish case, overall this turned out to get worse for the non-participants.

Several papers had previously looked at this experiment and concluded the intervention was a great success because participants fared so much better. But the result can of course not be generalized. What if everyone searches more for the same number of vacancies? Nothing changes much, except that vacancies may be filled faster. And what if the number of vacancies increased in those two counties because of the treatment, to the detriment of the other counties? Then applying the program to the whole country should not make a difference. Given the cost of these studies, this is a very disappointing result.

Is Africa doing much better than we thought?

When economic historians will look back at the global economy of last decade or two, they will likely summarize them as marked by a big recession in western economies, tremendous growth and convergence in Asia and South America, and stubborn lack of growth in Africa. As usual, one would say for Africa, which is a really frustrating continent.

Alwyn Young writes that this last assessment may be all wrong because the official statistics are biased downward. Looking at the consumption of durable like cell phones, cars, housing and health, as well as the use of time on the market by women (all reported by the Demographic and Health Survey), he finds that growth rates are more than triple what is indicated in the official statistics. How could they (or him) be so wrong? For one, price data is almost non-existent, which makes deflating nominal statistics rather hazardous. Second, it looks like the informal sector is being vastly underestimated. Mozambique and very recently Ghana have seen their GDP multiplied after the analysis of detailed surveys of their economies. If Young is right, Africa may not be quite catching up, but at least it is not losing ground.

How many dollars are abroad?

If you divide the amount of US dollars in circulation by the number of people living in the United States, you get an amount in the order of $3000. This is quite stunning, yet could be explained by various factors: money held in freezers and mattresses, money held by businesses, money used in the underground economy, lost money, and money held abroad. The latter is usually thought to make the bulk of it, given the status of the US dollar as an international reserve currency and its use as a parallel currency in several countries (if not full dollarization).

Edgar Feige thinks that in facts less than a quarter is abroad, but there is considerable uncertainty. For one, confidential data about dollar movements abroad has been made public, and official estimates need to be revised down. Also, indirect methods used to estimate this share seems to deeply flawed and very sensitive to their assumptions. But, somehow, it is still noticeable that the demand for dollars abroad declined with the emergence of the Euro as a viable alternative. One consequence of all this uncertainty and likely downward revision is that estimates on the size of the shadow economy that rely on money demand are wrong as well, something I actually complained about recently.

What drives the size of the shadow economy?

Why is the size of the informal sector larger in some countries than others? While it is easy to think what can drive off formal activities is easy, quantifying the contribution of each is difficult, foremost because we do not have much information about the informal sector, it is escaping the vigilance of the government, after all.

Friedrich Schneider and Andreas Buehn have a go at it. Schneider has made a living in computing and improving proxies for the size of the informal sector. They use his latest estimates for 39 OECD countries and regresses them on various indicators. While one may have doubts about the power of such estimates given the small sample lengths and the obvious uncertainty about the validity of the data, especially in a time series, one can still learn a few things, hopefully. In order of importance, these are the major factors for informality: indirect taxes, self-employment, unemployment, income tax and tax morale. But beyond doubts about sample size and degrees of freedom (what fixed factors are used in the panel regressions?), I am mostly concerned that the informality proxy used here (MIMIC) is constructed with some of the very variables (or close correlates) that are used in the regressions of the paper, namely the tax burden. I would have expected that you would use a different measure in this context, like a survey-based one or currency demand. Or used a different method than plain old OLS.

Put some economics back into spatial econometrics

One of the hot areas for econometric research in recent years has been spatial econometrics. Think of it, at least initially as time series econometrics in a different dimension. One interesting aspect of it is that instead of being single-dimensional like time series, it can be two-dimensional, or even more I guess. This field brings interesting new challenges, and it must be exciting working in this field. However, as too often in econometric theory, research becomes quickly detached from reality, and more specifically from the needs of empiricists. N never goes to infinity, for example.

Luisa Corrado and Bernard Fingleton bring forward another important point. These techniques are used to test economic theories, so one should be able to embed some restrictions from economic theory. It is all nice and sweet when one can find an optimal weighting matrix with the right properties, but it is useless if the found weights cannot be matched with anything one wants to test. The causality goes the wrong way: first determine restrictions from the theory, then use the constraints to find the optimal weighting matrix.

This is not just a theoretical consideration. Spatial lags are crucial in spatial econometrics and are suppose to capture some network effects. But they can also soak up the impact of latent or unobserved variables, as in "regular" econometrics. This can lead to severe miss-specification and biased inference, somethings one is all to familiar with using lags in time series. In fact, one should be downright suspicious of any time-series results that only holds when lagged dependent variables are used. The same must apply to spatial econometrics.

Risk preferences are heterogeneous across countries

In any international economic model I can think of, or any study comparing countries using economic models with utility maximizing utility, it is never considered that the preferences of households may differ. In closed economy models, some heterogeneity in preferences may be considered, but it is generally avoided because it is difficult to measure and in most cases would not make much a different anyway. But if we are thinking about some international imbalances, why not think about differences in aggregate preferences?

Marc Oliver Rieger, Mei Wang and Thorsten Hens look at the results of a survey administered across 45 countries that tries to elicit measurements about risk aversion, loss aversion and subjective probabilities. Too bad they did not consider discounting and the intertemporal elasticity of substitution. Anyway, they find that there are actually large differences across countries, differences that they attribute partially to economic conditions and "culture." For the former, the authors looks a GDP per capita and the human development index. I would also have looked at a measure of financial development. It seems to me that people become financially more sophisticated and thus willing to take risks if they are more exposed to financial markets. But I can be proven wrong.

A history of national accounting

We use national account data without realizing the efforts that lie behind the construction of these account. And frankly, it is sometimes like sausage making, but less and less so as the United Nations have managed to enforce rather widely its standards. But getting to these standards has been a very process.

Frits Bos documents the history of national accounting over three centuries. He shows that significant efforts and advances in national accounting arouse from two kinds of needs: understanding recent major crises and the increase of influence of the state and its policy making. To fix an economy, you need to know what is wrong with it, or at least the symptoms. To set policy, one needs measures to establish goals and how they have been reached.

How much tax evasion is there in the US?

As mentioned before, tax authorities are now especially eager to catch tax evaders. But how many of those are actually out there? One would suspect that there are relatively few of them in a low tax country like the United States. But again, the tax authority there has been given relatively few means to pursue investigations and audit rates are surprisingly low. And the tax code is so complicated that the line between tax evasion and confusion is rather blurred.

The Internal Revenue Service, the US federal tax authority, has estimates about how much it is missing in revenue, but as far as I know these numbers are kept well hidden, except for a study in the eighties. Academics have tried to replicate this exercise, obviously with poorer data than the IRS, but with less political pressure. The latest attempt is by Edgar Feige and Richard Cebula. They use a technique similar to one used to calculate the size of an informal economy, a technique based on the quantity of currency in circulation and of check deposits, adjusting for currency suspected abroad and financial innovation. Indeed, tax evaders try to hide income from reporting by using cash transactions. This ignores though those who use tax havens, and I welcome informed guesses on how large a factor that may be.

In any case, Feige and Cebula come to the conclusion that about 20% of reportable income is not properly reported, leading to lost revenue in the order of $400-500 billion every year. They even compute a time series that allows them to figure out what makes the non-compliance rate change. It will not surprise that it increases when national income is higher, when tax rates are higher or when nominal interest rates are higher. It is interesting to see that higher unemployment rates lead to higher non-compliance. That may have to do with more people getting informal income while on unemployment insurance. Still, I cannot shake the feeling that all these results are shaky themselves, as this data is essentially made up.